ML Engineer, Character Models – Diffusion

This is a fully remote position, open to applicants in Canada, +1 more country.

📋 Description

• Train and optimize diffusion models, pipelines, and adapters for character development.

• Enhance the evaluation framework to assess character fidelity in conjunction with overall output quality.

• Guarantee that a child's verbal request for a character consistently yields the correct outputs.

• Collaborate with brand teams to secure approvals and facilitate product launches.

• Work alongside the ML Serving engineer to transition new products into production.

• Team up with the Trust & Safety engineer to ensure products are safe for children from the outset.

• Develop and refine models and pipelines for generating approved, on-brand character stickers for Stickerbox.

• Establish the data and evaluation systems that support the models.

⛳️ Requirements

• Practical experience in fine-tuning diffusion or other generative image models.

• Proficiency in constructing training data pipelines with genuine provenance requirements.

• Experience in developing evaluations for generative output and utilizing them for decision-making.

• Strong programming skills in Python.

• A curious mindset and a willingness to innovate and contribute to the future of play.

🏝️ Benefits

• Opportunities for growth and career advancement.

• Flexible working hours.

• A culture of innovation, collaboration, and enjoyment.

• Comprehensive benefits package.

• Additional enjoyable perks!

• Reasonable accommodations provided during the hiring process.